www.nature.com/scientificreports
OPEN Sarcophyton glaucum *
Certain sof corals and all reef-building corals form associations (collectively termed “holobionts”) with symbiotic dinofagellates and bacteria; these microbes are critical to the physiological function and survival of their animal hosts1−3. Although the mutualistic dinofagellates translocate photosynthetically fxed carbon into host cytoplasm, corals and other endosymbiotic anthozoans nevertheless rely on heterotrophy, as well, for nourishment 4−5. Tese mutualistic associations consequently span a bridge between benthic and planktonic food webs and consequently play critical roles as primary and secondary producers 6. Te culture of alcyonacean sof corals (Cnidaria: Anthozoa: Octocorallia: Alcyonacea)- notably Sarcophyton, Sinularia, and Lobophyton- has advanced rapidly in recent decades given the needs to (1) explore the impacts of environmental change on their physiological performance and (2) develop a sustainable supply for both the marine cosmeceutical and aquarium trade industries1,7,8. Many cultured sof corals used in natural product research rely on natural seawater fow-through systems (FTS)9,10; however, natural product yields may difer from those of wild populations11. Other studies have used recirculating aquaculture systems (RAS) in which seawater quality parameters can be readily and automatically modulated by a variety of microprocessor-based dosing systems12−14. While biological factors, such as “live rocks” (typically dead coral skeleton encrusted by crustose coralline algae) that accumulate rich microbial fora15,16 and heterotrophic feeding4−5,17, have been explored more recently, a systemic comparison across diferent types of culture systems has not yet been achieved for sof corals. Light intensity and fow velocity have been examined in numerous sof coral experiments 7,17−19. Te former is especially important for sof corals harboring dinofagellates, whose photosynthetic performance can directly or indirectly afect the physiology and growth of their hosts 20,21. Light has also been shown to afect not only growth, but also secondary metabolite production, in symbiotic sof corals 19. With respect to fow, sof corals
National Museum of Atlantic Oceanographic and Meteorological Laboratory, Cooperative Institute for Marine and *email: [email protected]
| (2020) 10:20200 | www.nature.com/scientificreports/
Figure 1. Schematic of experimental design alongside key fndings. An image of sof corals mounted to ceramic pedestals and the tank system (A). Representative pedestals and sof coral fragments have been shown from the RAS−B (B), RAS+B (C), and FTS (D).
alter their morphologies in response to changes in seawater velocity 18, and seawater fow promotes material exchange and metabolism (thereby afecting metabolite yield). However, only single-factor experiments have been conducted with sof corals, despite the well-documented interaction between irradiance and water fow in hard corals like Galaxea fascicularis22. Tus, research focused on the interaction of light and fow on sof coral physiology is needed. Measurements of morphology and growth in sof corals are more difcult compared to stony corals23; the latter consist of a large amount of calcium carbonate (exoskeleton) covered by a thin veneer of live tissue. Alcyonacean sof corals, in contrast, are mostly feshy, with only small amounts of calcium carbonate (sclerite) skeleton. Shape is instead maintained by hydrostatic pressure from water pumped into a canal system. Such hydroskeletons change quickly in size and shape in response to environmental shifs. Terefore, it is difcult to accurately measure their size and growth 23, though oral disc diameter (ODD), stalk diameter, colony height, and wet and dry colony weight have all been assessed in prior works 24,25. In general, ash weight and ash-free dry weight (AFDW) are thought to represent skeleton and organic matter content, respectively 25. Sof corals of the genus Sarcophyton are abundant on many coral reefs in the Indo-Pacifc Ocean, and par- ticularly Taiwan 6. Sarcophyton colonies are mushroom-shaped, with (1) feshy stalks elevating the polyp-bearing oral discs and (2) sclerites in the interior feshy tissues of the colony. Sarcophyton glaucum (Quoy & Gaimard, 1833) is a suspension feeder that lives in mutualistic association with phototrophic dinofagellates (family Sym- biodiniaceae), as well as an array of other microbes. It has been cultured in FTS to investigate natural product production9 and in RAS to assess the efects of light intensity and fow velocity on its physiology17,20,26,27. Given the recent recommendation to develop the capacity to biobank and biopreserve corals ex situ 28, a rigorous com- parison of sof coral performance across diferent culture systems is of need. Herein we aimed to study the efect of three diferent types of culture systems—a RAS with no microbial fora or exogenously supplied food [RAS minus (−) additional biological input = RAS−B], a RAS without a protein skimmer supplemented with live rocks and a supplemental phytoplankton solution fed corals [RAS plus ( +) additional biological input = RAS+B], and a traditional FTS- on the physiological performance of S. glaucum (Fig. 1). Within each culture system, two levels of photosynthetically active radiation (PAR; 100 or 200 μmol quanta m−2 s−1) and fow velocity (5 or 15 cm s−1) were administered, and several coral physiological response variables were measured to assess performance: colony color, height, base diameter, growth rate, ODD, and percent (%) organic weight (i.e., AFDW). We hypothesized that the physiological performance of the corals would be superior in the RAS+B system due to the (1) potential for heterotrophy and (2) the fact that feeding was carried out in separate feeding tanks (which would potentially limit macroalgal growth in the culture tanks). Te pH, concentrations of Ca2+and Mg2+, and carbonate hardness/alkalinity (KH) were sig- nifcantly higher in RAS−B and RAS+B than in FTS (Table 1). Levels of ammonia, nitrite, nitrate, and phosphate in the three systems remained below detectable levels (< 0.2 mg L−1) during the entire experimental period (data
| (2020) 10:20200 | Vol:.(1234567890) www.nature.com/scientificreports/
Culture system Temperature (°C ) Salinity pH Ca2+ (mg L−1) Mg2+ (mg L−1) KH (dKH) RAS−B 25.7 ± 0.10B 34.8 ± 0.10 8.15 ± 0.03A 450 ± 4.80A 1318 ± 6.32AB 7.4 ± 0.15A RAS+B 26.0 ± 0.11B 34.9 ± 0.10 8.11 ± 0.03AB 447 ± 5.27A 1320 ± 5.00A 7.3 ± 0.17A FTS 26.5 ± 0.08A 35.0 ± 0.10 8.03 ± 0.02B 411 ± 4.55B 1297 ± 4.41B 7.0 ± 0.13B
Table 1. Comparison of seawater chemistry parameters across the three culture systems: RAS−B (n = 13 measurements), RAS+B (n = 9), and FTS (n = 9). When the non-parametric ANOVAs detected diferences across the culture systems, Dunn’s multiple comparisons tests were carried out between individual means, and signifcant diferences (p < 0.05) are denoted by capital letters. All error terms represent standard error of the mean.
Figure 2. Violin plots depicting temporal variation in buoyant weight (A), oral disc diameter (ODD: B), and fragment color (C) of sof corals over time in the three culture systems—recirculating aquarium system (RAS) without additional biological input (RAS−B; red), RAS with exogenous food supply and “live” rocks (RAS+B; green), and a fow-through system (FTS) featuring partially fltered seawater (blue)—as well as interaction plots of raw color data (D). It should be re-emphasized that RAS−B tanks were converted to RAS+B ones afer day-84. Error bars spanning mean values at each sampling time represent standard error, asterisks in (D) denote Tukey’s honestly signifcant diferences (p < 0.05) between RAS+B and FTS corals, and lowercase letters in (D) denote temporal diferences (Tukey’s HSD, p < 0.05) in color for corals of the RAS−B culture system. RAS+B vs. FTS diferences in (D) are instead denoted by asterisks (*).
not shown). PAR and fow did not deviate signifcantly over time (p > 0.10) and remained close to their target values (see details in “Methods”).
Although all fragments in the three culture systems were mushroom-shaped (Fig. 1B–D), many sof coral fragments cultured in the RAS−B system appeared unhealthy (Fig. 1B); polyps were not always extended, and neither their buoyant weight [BW; Fig. 2A; one-way ANOVA efect of time (for this and all following tests in this paragraph), F = 1.77, p = 0.134] nor their ODD (Fig. 2B; F = 0.0629, p = 0.8023) increased signifcantly over time. Tey also paled signifcantly over this period when pooling across all light × fow treatments (Fig. 2C; F = 53.1, p < 0.0001), though the color decrease of RAS−B corals of the two low-light treatments [high-light + low fow (Fig. 2D-1) and high-light + high fow (Fig. 2D-2)] was not statistically signif- cant. Afer changing to the RAS+B system (see Fig. 1C for representative fragment images.), their ODD enlarged (Fig. 2B; F = 9.54, p < 0.01), and their pigmentation increased (Fig. 2C; F = 32.8, p < 0.0001). Green, flamentous
| (2020) 10:20200 | Vol.:(0123456789) www.nature.com/scientificreports/
Figure 3. Principal components analysis of correlations across standardized sof coral physiological response variable data. Please note that specifc growth rate (SGR), color change, and oral disc diameter (ODD) all refect rates or changes over time (day−1, fnal–initial, and % change day−1, respectively); fnal values were instead incorporated for percent (%) organic weight (i.e., AFDW), which was only measured once. Tere was a clear inverse relationship between % ODD change day -1 and AFDW.
algae tended to overgrow the pedestals in the FTS (Fig. 1D), and neither buoyant weight (Fig. 2A; p = 0.635) nor ODD (Fig. 2B; p = 0.120) increased signifcantly between culture days 110 and 170. FTS fragment color actually decreased over this same period (Fig. 2C), though the global temporal diference was not statistically signifcant (p = 0.204; see also Fig. 2D for temporal changes for each of the four light × fow interaction groups.). In the principal components analysis (PCA) on correlations (Fig. 3), the frst two axes captured nearly 60% of the variation in the dataset, and the primary coordinate (x) axis featured ODD change (eigenvalue = 0.63) and color change (0.55) as the dominant positive loading factors, with AFDW as the dominant negative one (− 0.53). Te secondary (y) axis was primarily defned by the negative relationship between the SGR (0.44) and the color change (− 0.48). Tere was evident separation between RAS−B and RAS+B samples in the PCA, and a multivariate ANOVA (MANOVA) + canonical correlation analysis (CCA) verifed this (Wilks’ lambda of culture treatment efect = 16.6, p < 0.001). A more detailed explanation on how the individual treatments (light × fow) afected these sof coral response variables can be found in Fig. 2D, Table 2 (statistically signifcant fndings only), the online supplemental results (OSR), Supplementary Table S1 (the full repeated-measures ANOVA model), and Supplementary Figs. S1 and S2. Briefy, there were more signifcant efects of light on response variables measured in corals of the RAS−B [e.g., ODD growth (Supplementary Fig. S2G) and color change (Fig. 2D, Supplementary Fig. S2M)]. Flow more signifcantly afected the response of sof corals cultured in the FTS, particularly the AFDW (Supplementary Fig. S2L), which was higher at low fow rates. In contrast, neither light nor fow afected the SGR (Supplementary Fig. S2E) or ODD growth (Supplementary Fig. S2H) of RAS+B corals. As an excep- tion, the color increase was higher for RAS+B corals cultured at the higher light level (Supplementary Fig. S2N), though no signifcant post-hoc diferences were noted. All sof corals in the three culture systems survived, even those cultured for 170 days; this actually represents the longest experimental culture of sof corals (only 30 days–5 months in other studies14,20). However, corals grew slowly, and tended to become paler, in the RAS−B system; upon feeding (RAS−B modifed to RAS+B), they regained pigmentation, and their ODD enlarged. Furthermore, unlike the ceramic pedestals of the FTS, those of the RAS+B did not regularly become biofouled by algae. Tis may not only be due to the use of synthetic seawater, in which nutrients were not present, but also because RAS+B coral feeding was undertaken in a separate tank. We therefore avoided the eutrophication associated with feeding experiments carried out in RAS 12, in which algae may bloom and smother corals; when other studies carried out feeding directly in the RAS 17,26, such high nutrient loads were generated that sof coral growth was ultimately thwarted. We therefore recommend to those seeking to culture sof corals over long-term timescales to use RAS and feed their organisms in separate tanks. Fed corals demonstrate lower SGR than starved ones17. Herein the SGR of the fed corals in RAS+B was lower than in RAS−B corals for one (HLLF) of the four light × fow interaction groups; however, RAS+B ODD increased more rapidly than in the other two culture systems. Te fact that the % increase in ODD (but not AFDW) was
| (2020) 10:20200 | Vol:.(1234567890) www.nature.com/scientificreports/
Response variable efect df Exact F p Buoyant weight (increase day−1)a Culture system (see Fig. 2A for raw data.) 2 49.1 < 0.0001 Culture system × light 2 6.30 0.00330 Culture system × fow 2 4.36 0.0173 Colony height (cm; fnal value only) (Supplementary Fig. S1A–D, S2B,C) Flow 1 63.01 < 0.0001 Culture system × fow 1 7.60 0.0065 Culture system × light × fow 1 7.44 0.0071 SGR (day−1)b (Supplementary Figs. S1E–H, S2D–F) Culture system 2 55.6 < 0.0001 Culture system × fow 2 5.33 0.0075 Light × fow 1 10.20 0.0023 ODD (% change day−1) [Supplementary Figs. S1I–L, S2G–I, and Fig. 2B (raw data)] Culture system 2 7.12 0.0017 Culture system × light 2 3.61 0.0334 Percent organic weight (AFDW)b (Supplementary Figs. S1M–P, S2J–L) Culture system 2 14.2 < 0.0001 Color change (fnal-initial) (Supplementary Figs. S1Q–T, S2M–Oc) Culture system 2 43.5 < 0.0001 Culture system × light 2 13.7 < 0.0001
Table 2. Select two-way, repeated measures ANOVA results for the efects of culture system, light, fow, and their interaction(s) on several sof coral response variables. Only statistically signifcant (p < 0.01) fndings have been included; non-signifcant results, as well as Tukey’s honestly signifcant diferences and tank efects, can instead be found in Supplementary Table S1. Colony height (cm) was measured in corals of the RAS+B and FTS only (fnal sampling time only); since these two culture systems were independent, a standard 3-way ANOVA (culture system × light × fow) was instead used. No treatment factor afected base diameter (mm; see Supplementary Table S1 and Supplementary Figs. S1A–D; S2B,C.). Please note that, in Supplementary Figs. S1 and S2, individual one-way ANOVAs for determining the efects of culture system (RAS−B vs. RAS+B vs. FTS) and treatment (the four light × fow interaction groups), respectively, within each treatment and culture system, respectively, were instead carried out. AFDW = ash-free dry weight. ODD = oral disc diameter. SGR = specifc growth rate. a square root-transformed data. blog-transformed data. csee Fig. 2D for raw color scores over time for the 12 culture system × treatment groups.
higher in RAS+B corals is likely due to heterotrophy7. For instance, fed colonies of the symbiotic temperate coral Astrangia poculata exhibited signifcantly greater photosynthetic efciency than starved conspecifcs29. It should be noted that SGR and ODD data gave conficting results: RAS−B corals grew faster when looking at the former response variable, with ODD change revealing that RAS+B corals grew faster. Given the issues raised in the Introduction with using the BW technique with predominantly sof-bodied organisms, we argue herein that sof coral biologists carefully consider its validity; perhaps the easily measured ODD is a superior, alterna- tive metric for sof coral growth23,24,26. Indeed, others have suggested that BW data may be spurious since sof coral tissues contain a signifcant amount of water, meaning that the density of some colonies may be similar to that of pure seawater 25. Regardless, the SGR documented herein were comparable to, or even higher than, those reported in other studies of sof corals (Table 3), despite their very small starting sizes. Te average annual ODD growth rate was 1.13 cm year−1 in the three culture systems, similar to the linear growth rate of 1.0 cm year−1 of Sarcophyton (0.5–4.9 cm colony diameter) on the Great Barrier Reef 24. Although our growth rates were higher than those reported in other aquarium studies, it is worth noting that, when comparing husbandry to in situ data, sof corals grew faster on the reef in the lone study that made this comparison 26; it will be critical, then, to gather feld data from our Taiwanese feld sites to ensure that aquarium growth rates are comparable to in situ ones. Unlike RAS−B and FTS corals, RAS+B specimens darkened over the duration of the experiment; this refects either an increase in Symbiodiniaceae density or pigmentation within their cells (or both) and may indicate that starved endosymbionts of the RAS−B in particular were nitrogen limited (especially given the virtual absence of nitrates and nitrites in the seawater)5. However, whether the presumed infux of nitrogen and other nutrients occurred via feeding on the exogenously supplied phytoplankton feed (direct feeding)4, feeding on the microbial fora released from the live rocks (indirect feeding), or some other mechanism remains to be determined. It is worth noting that the presence of live rock was associated with higher coral survival rates and Fv/Fm values, as well as fewer bleached specimens in hard coral studies 15, and we advocate continued research on the role of microbiology in marine animal health and husbandry 1,3,7,8. Organic weight (AFDW) is a reliable metric for normalizing sof coral data25, particularly when inter-species comparisons are made, and it is thought to better refect the proportion of feshy coral tissue. AFDW was only signifcantly afected by fow rate in the FTS, and percentages documented herein (26–33%) were similar to those
| (2020) 10:20200 | Vol.:(0123456789) www.nature.com/scientificreports/
Species Culture system SGR (day−1 × 100) Size BW (g) Reference Sarcophyton glaucum RAS−B 0.069–2.765 11–13 mm 0.243–0.434 Herein Sarcophyton glaucum RAS+B 0.025–1.072 12–15 mm 0.295–0.526 Herein Sarcophyton glaucum FTS 0–1.828 13–14 mm 0.367–0.647 Herein Sarcophyton cf. glaucum RAS+B 0.027–0.028 ~ 30 mm 27 Sarcophyton cf. glaucum RAS+B 0.035–0.04 ~ 40 mm 20 Sarcophyton cf. glaucum RAS+B 0.11–0.39a 1.5 cm2 17 Sarcophyton spp. RAS+B 0.055–0.380 0.69–0.782 14 Sarcophyton glaucum RAS+B 6 mm 0.0077c 26 Sinularia fexibilis RAS+B 0.039–0.043 10 cm 21 Sinularia fexibilis RAS 0–0.016b 5–7 cm 0.014–0.033a 19 Sinularia fexibilis RAS 0.008–0.019b 5–6 cm 18
Table 3. Comparison of study species, culture system, specifc growth rate (SGR), size, and buoyant weight (BW) of coral fragments in sof coral culture studies. Te size data for Sarcophyton and Sinularia fexibilis have been presented as diameter (or area in cm2) and length, respectively. a Values were estimated from the fgures. bAverage weekly value. cDry weight.
reported for congenerics in situ (29%)24 and higher than those of cultured conspecifcs (20–29%)17. On the other hand, S. glaucum values showed much larger variation (~ 10 to 55%) in a prior work that also compared diferent culture conditions and systems 26. Light intensity afected the ODD in the RAS−B (100 > 200 μmol quanta m −2 s−1), and both the ODD and AFDW were afected by fow in the FTS. Water fow reduces the difusive boundary layer around the coral and infuences the supply and uptake of dissolved gasses, nutrients, and food, as well as the removal of sediments and metabolic waste products; collectively, then, higher fow rates should beneft the health and growth of corals 22, and such has been demonstrated in symbiotic sof corals 24. Herein, though, corals generally grew faster at the lower of the two fow rates; perhaps exposure to constantly high fow rates of 15 cm s−1 results in stress to the polyps or actually sweeps food past too quickly for particles to be efectively captured. Light more signifcantly afected growth (as ODD) in corals of the RAS−B. Tis is likely because these corals were dependent on autotrophy alone for nutrition. In contrast, there were few efects of light intensity or fow velocity in corals of the RAS+B, potentially suggesting that exogenous food supply was sufcient for coral growth; had it not been, the corals may have been more sensitive to changes in their abiotic environment. In the future, it will be worthwhile to determine whether fed sof corals obtain relatively more energy and nutrition from heterotrophy than autotrophy; it can only be stated with the data in hand that fed corals generally outperformed starved ones from assessment of the response variables measured. We demonstrated that light and fow efects on sof coral physiology are culture system-dependent. Given that (1) sof corals are mostly feshly and (2) ODD measurements can be made in only several seconds (versus minutes for buoyant weighing), we advocate assessing both ODD and the BW-based SGR in future studies of sof corals, especially Sarcophyton. Sof corals cultured in RAS+B featuring live rocks and an exogenous food supply tended to grow more quickly and presented darker pigmentation. Furthermore, by feeding these corals in a separate tank, macroalgal biofouling was virtually eliminated from the RAS+B husbandry tanks. We therefore recom- mend the future culture of sof corals in RAS+B systems and advocate that similar such experimental marine animal husbandry approaches be conducted elsewhere (and with additional species), especially given the recently identifed need to optimize aquarium husbandry for biobanking/biopreserving marine organisms whose habitats have been compromised by climate change and other anthropogenic stressors 28. Tree S. glaucum colonies (ODD = ~ 30 cm) were collected under Kenting National Park permit 1570001572 (to TYF) at depths of 7–8 m outside the inlet of Taiwan’s third nuclear power plant (21° 57 15.7" N, 120° 45 21.2" E) in Nanwan Bay, Southern Taiwan. Colonies, which were at least 4–5 m apart, were iden- tifed in situ by assessment of their morphological characteristics6, quarantined in the husbandry facility of the National Museum of Marine Biology and Aquarium for a week, and acclimated in a 200-L fow-through tank char- acterized by the following conditions: natural seawater fltered to 5 μm, temperature = 26 ± 1 °C (mean ± standard error for this and all other error terms unless stated otherwise), salinity = 35 ± 1, and PAR = 100 ± 1 μmol quanta m−2 s−1 (light–dark cycle of 12 h/12 h). Afer this initial acclimation period, a sterilized scalpel was used to cut fragments (~ 10 mm in diameter and ~ 0.3 g BW) along the edge of the oral disc. Each of the three parent colonies produced 81 fragments, and absorbable polyglycolic acid stitches (VISORB, 1/0)30 were used to attach them to 2.7-cm, etched, T-shaped ceramic pedestals (average BW = 5.167 g); these stitches have been shown to be associated with shorter recovery and attachment times than the more commonly used rubber bands 14,20. Te 243 fragments were placed in a 200-L tank under the same conditions as above for recovery and attachment for 2 weeks, and all survived the preparatory processes (Fig. 1A). Of these, 108 were used in the RAS−B, and 72 of these were carried over into the RAS+B. A separate 108 were used in the FTS. All three culture systems
| (2020) 10:20200 | Vol:.(1234567890) www.nature.com/scientificreports/
are described in detail below. Since 72 of the 216 total fragments were analyzed in a repeated measures fashion (RAS−B modifed to RAS+B), a total of 288 measurements were made for the non-destructive response variables described below.
RAS−B. Te RAS−B included synthetic seawater: Red Sea salt (Red Sea Aquatics, Ltd.) mixed with reverse osmosis (RO) water (no additional microbial fora). In order to ensure consistent water qual- ity, all six experimental tanks (each 60 × 35 × 20 cm; Fig. 1A) were connected in series to a 240-L “life support” tank (120 × 45 × 45 cm), which contained a 0.2-mm flter bag, an automatic Mato-2009 RO bucket (Autoaqua, Taiwan), a protein skimmer (JNS, CO2, Taiwan), a zeolite drum (JNS, ZR-2), a primary pump (Mr. Aqua, 6000 L/H, Taiwan), a titration system (Johnlen, CS072A-1, Taiwan; for measuring alkalinity [as KH] and concentra- tions of Ca2+ nd Mg2+), a heater (ISTA, 350W, USA), and a chiller (Resun, C-1000 p, China; 26 ± 1 °C). Te salinity was maintained at 35 using a Mato-100P osmoregulator (Autoaqua) that automatically compensated for evaporative water loss by adding fresh RO water. At the beginning of the experiment, 108 fragments were randomly placed across the 6 experimental tanks (n = 18 fragments/tank), where they were then cultured for 84 days. At the end of this period, 36 fragments were randomly selected and stored in a − 20 °C freezer for analy- sis of organic matter (AFDW; described below).
RAS+B. Te remaining 72 fragments were lef in the six tanks, and the protein skimmer was removed to keep dissolved and particulate organic material in the synthetic seawater column; it was hypothesized that doing so would enhance the biological activity in the system and potentially beneft the remaining corals 12. Live rocks (50 kg) were also added because, similarly, it was hypothesized that doing so could increase the growth and color of the sof corals (color paled over the frst 84 days in the RAS−B treatment; see Fig. 2B,C). Te addition of live rock, an efective bioflter, can improve nutrient cycling and shif microbial communities towards a more typical seawater assemblage15−16. Tus, the RAS−B was modifed to a RAS+B, and the 60-day RAS+B vs. FTS (see below) experiment began on the 110th day. In addition to the features listed above, corals of the RAS+B culture system were fed with a reef phytoplankton solution (Seachem, USA) containing concentrated microalgae (Talassiosira weissfogii, Isochrysis sp., & Nan- nochloropsis sp.) once every 3 days. Afer the lights were turned of, fragments were moved into an independent, 20-L feeding tank (60 × 35 × 20 cm) with bubble stones in the four corners (to allow for even water mixing) and a heater that maintained the temperature at 26 °C. Afer 30 min, the polyps and tentacles extended, and 3 mL of the feeding solution were added. Afer 4 h, the fragments were removed from the feeding tank, rinsed with fltered seawater, and returned to the experimental tanks. It was hypothesized that, by feeding in a separate tank, the elevation in nutrient levels and macroalgal biomass associated with the feeding process would be avoided.
FTS. For these tanks (n = 6), natural seawater was frst pumped in from the nearby ocean (Houwan Bay), fl- tered through sand (50 μm) and passed through 30-ton coral reef mesocosm tanks (described previously 31) in which seawater temperature was maintained at 26 ± 1 °C. Afer fltering sequentially (100, 50, 5 μm), the seawater then entered the experimental tanks. Te water fow-through (tank volume h −1) was 10 L h −1, and there was no “life support” tank. Te 108 coral fragments were transferred from the acclimation tanks (where they had accli- mated for 110 days) to the FTS tanks at the same time as the RAS−B were converted to RAS+B, and the FTS vs. RAS+B comparison was carried out over 60 days (culture days 110–117; see Fig. 2).
Both the RAS−B and RAS+B systems were connected to a 240-L life support tank (315 L in total), and partial synthetic seawater in the RAS−B and RAS+B was changed weekly (30 L; ~ 10% of the volume of the entire culture system). For all three culture systems, concentrations of nutrients (nitrate, nitrite, phos- phate, and ammonia), calcium (Ca 2+), magnesium (Mg2+), carbonate hardness/alkalinity (KH), and pH of the three systems were measured weekly (Salifert Prof Test, Holland).
A PVC plate was placed in the center of each of the six tanks within each of the three culture systems (Fig. 1A), and LED lights (Illumagic, ComboRay G2, Taiwan) were programmed to administer 100 μmol quanta m −2 s−1 (low light) to one half of each divided tank and 200 μmol quanta m −2 s−1 to the other (n = 36 experimental areas across the 18 tanks, each at a 12-h/12-h light/dark cycle). Light levels were chosen based on levels used in prior studies that were associated with relatively high sof coral growth rates17,19−21,26−27. Each experimental tank also contained a fow motor (Maxspect, GP-03, China), and half of the six tanks were exposed to a fow of 5 cm s−1, with the remaining three at a higher fow of 15 cm s−1. As for the PAR levels employed, these fow rates were set based on those that were associated with relatively high sof coral growth in prior studies17,19−21,26−27. Tus, there were four treatments in triplicate within each of three culture systems: high light/low fow (HLLF), high light/high fow (HLHF), low light/low fow (LLLF), and low light/high fow (LLHF). Te low/ high light intensity averaged 105 ± 0.15/206 ± 0.25 μmol quanta m −2 s−1, and the low/high fow velocity averaged 4.6 ± 0.06/15.7 ± 0.16 cm s−1, as measured by light (Li-Cor, LI-193SA, USA) and fow (Kenek, GR20/GR3T-2- 20 N, South Korea) meters, respectively. For RAS−B and RAS+B, 9 and 6 sof coral fragments were analyzed in each of the 12 experimental areas (2 areas for each of the 6 experimental tanks), resulting in a total of 108 and 72 fragments for the light × fow experiment, respectively. For the FTS, 9 fragments were used, resulting in a total of 108 fragments.
| (2020) 10:20200 | Vol.:(0123456789) www.nature.com/scientificreports/
Te weights of the coral fragments were measured by a buoyant weighting technique on a Mettler Toledo AB204 balance (precision = 0.1 mg; USA). A glass beaker containing fltered seawater (26 ± 0.5 °C and salinity of 35) and a thermostatic bath were placed under the electronic scale, and the coral fragments were suspended on fshing line under the electronic scale for buoyant weight measurements. Before each measure- ment, the surface of the coral pedestal was lightly brushed with a toothbrush to remove algae. RAS−B fragment weights were measured every 3 weeks, while RAS+B and FTS fragments were weighed biweekly. Te SGR (day −1) was calculated as: (lnWf − lnWi)/Δt, where lnWi and lnWf represented the natural logarithms of the coral frag- ment buoyant weights (technically buoyant mass, g, but regularly referred to as “weight” in virtually all publica- tions) at the beginning and the end of the experiment, respectively, and ∆t represented the duration in days. We generally multiplied these data by 100 in most fgures. Although we have shown the raw BW data for each culture system (Fig. 2A), we generally prioritized the SGR in our discussion since, unlike raw BW data, it accounts for potential starting size diferences across corals.
Te mean diameters ((length + width)/2) of the oral discs were measured at night (when tentacles had retracted) with vernier calipers (readable to 1 mm) at times = 0 and 84 days for RAS−B. A % change in ODD over this time period was calculated, since, as with BW, a relative rate of change theoretically better accounts for start- ing size diferences across sof corals [as well as the fact that culture durations difered between RAS−B (84 days) and RAS+B and FTS (70 days)]; for this reason, temporal diferences in raw ODD across all four treatments have not been depicted for each culture system (data pooled across treatments for each culture system can instead be seen in Fig. 2B, with rates of change found in all following fgures). For FTS and RAS+B, measurements were made on culture days 110 and 170, with the % change in ODD also calculated. ODD % change day −1 has gener- ally been shortened to “ODD growth” throughout the article, though, for some correlation-based analyses (see below.) raw ODD values were instead used.
Te height and mean base diameter ((length + width)/2) of each colony were measured at night when the polyps were retracted with vernier calipers (as above) on culture day 170 (RAS+B and FTS only).
Coral fragments (n = 36, 72, nd 108 for RAS−B, RAS+B, and FTS, respectively) were removed from the pedestals and dried in a vacuum freeze dryer (Xian Toption Lyophilizer, China) for 48 h. Te dry weight was measured, afer which the samples were incinerated at 450 °C in an MF-30 mufe furnace (Hipoint, Taiwan) for 2 h to calculate the inorganic weight. AFDW was calculated (as a proxy for % organic weight) as (dry weight- inorganic weight)/dry weight × 100.
Coral fragments were photographed with a fxed light source (5500 K, LED) in a 40 × 40 × 40 cm studio at the beginning and end of the experiment using a digital camera (Olympus, Tough TG-5). Based on CoralWatch’s “Coral Health Chart”32, fragments were scored along the E1 to E6 axis, and the changes in color scores (fnal assessment level-initial), rather than color itself, were assessed in the statistical tests outlined below to eliminate bias due to certain corals beginning experiments at slightly paler levels than others. Tat being said, we did plot raw color values over time for each treatment × culture system interaction group (Fig. 2D) to high- light diferential efects of treatment on color changes for RAS+B and FTS corals, since, in some cases, starting pigmentation levels were similar.
Data were frst tested for normality (Shapiro–Wilk’s test of residuals) and equal vari- ance (Levene’s test), and, since seawater quality parameters did not conform to these assumptions, they were analyzed by non-parametric Kruskal–Wallis tests (on ranks) to determine the efect of culture system on each parameter. Dunn’s multiple comparisons post-hoc tests were performed to compare individual culture system means. Because the RAS−B and RAS+B experimental stages were not independent, they were treated as repeated measures (two-way) in the statistical analysis of the physiological response variables for (1) BW (raw values and raw increases per day), (2) SGR (% change day−1), (3) % change in ODD day−1, (4) organic weight (AFDW; fnal time only), and (5) color. For these analyses, both the temporal change (excluding AFDW) and the fnal values were assessed separately. For those two parameters (colony height and base diameter) assessed only in corals of the latter two stages, RAS+B and FTS, the more traditional 3-way ANOVA (culture system × light × fow) was instead carried out since these two culture systems were fully independent. For two-way, repeated measures ANOVAs, tank was nested within light (df = 1) × fow (df = 1), whereas for three-way ANOVA, tank was nested within culture condition (df = 1 since RAS−B data were excluded from these analyses) × light × fow. Tukey’s honestly signifcant diference (HSD) post-hoc tests were used to compare individual mean diferences among culture systems, across treatments, or in response to light and/or fow. As a less conservative approach, one-way ANOVAs were carried out to test for the efect of culture system for each of the four light × fow treatments (Supplementary Fig. S1) for each of the following response variables: colony height, colony base diameter, SGR, ODD (fnal values and % changes), organic weight (fnal values only), and color (fnal values and raw changes). Similarly, two-way ANOVAs (light × fow) were carried out within each culture system for the same response variables (Supplementary Fig. S2), and one-way repeated measures ANOVAs were used to evaluate the efects of time and culture system on BW, ODD, and color (Fig. 2). Alpha levels of 0.01 and 0.05 were set a priori for ANOVAs and post-hoc tests, respectively. To depict variation across culture systems within a multivariate framework, a principal components analysis (PCA) on correlations was carried out with standardized data from the following response variables: SGR (%
| (2020) 10:20200 | Vol:.(1234567890) www.nature.com/scientificreports/
day−1), ODD (% change day−1), organic weight (%; fnal values only), and change in color (fnal-initial values). A canonical correlation analysis (CCA) based on a multivariate ANOVA (MANOVA) was also carried out to uncover multivariate diferences between culture systems, and Wilks’ lambda was calculated to assess statistical signifcance. All statistical analyses were performed using JMP (ver. 14.2).
Animals involved in the experiments are not listed in CITES and were cultured in the laboratory for experimental purposes only.
Received: 25 June 2020; Accepted: 29 October 2020
1. Leal, M. C. et al. Coral aquaculture to support drug discovery. Trends Biotechnol. 31, 555–561 (2013). 2. Peixoto, R. S., Rosado, P. M., Leite, D. C. A., Rosado, A. S. & Bourne, D. G. Benefcial microorganisms for corals (BMC): Proposed mechanisms for coral health and resilience. Front. Microbiol. 8, 341 (2017). 3. van de Water, J. A. J. M., Allemand, D. & Ferrier-Pagès, C. Host-microbe interactions in octocoral holobionts—recent advances and perspectives. Microbiome 6, 64 (2018). 4. Piccinetti, C. C. et al. Herbivory in the sof coral Sinularia fexibilis (Alcyoniidae). Sci. Rep. 6, 22679 (2016). 5. Houlbrèque, F. & Ferrier-Pagès, C. Heterotrophy in tropical scleractinian corals. Biol. Rev. 84, 1–17 (2009). 6. Benayahu, Y., Jeng, M. S., Perkol-Finkel, S. & Dai, C. F. Sof corals (Octocorallia: Alcyonacea) from southern Taiwan. II. species diversity and distributional patterns. Zool. Stud. 43, 548–560 (2004). 7. Leal, M. C., Ferrier-Pagès, C., Petersen, D. & Osinga, R. Coral aquaculture: Applying scientifc knowledge to ex situ production. Rev. Aquacult. 6, 1–18 (2014). 8. Leal, M. C. et al. Marine microorganism-invertebrate assemblages: Perspectives to solve the “supply problem” in the initial steps of drug discovery. Mar. Drugs 12, 3929–3952 (2014). 9. Huang, C. Y. et al. Glaucumolides A and B, biscembranoids with new structural type from a cultured sof coral Sarcophyton glau- cum. Sci. Rep. 5, 15624 (2015). 10. Sang, V. T. et al. Coral and coral-associated microorganisms: A prolifc source of potential bioactive natural products. Mar. Drugs 17, 468 (2019). 11. Peng, B. R. et al. Aquaculture sof coral Lobophytum crassum as a producer of anti-proliferative cembranoids. Mar. Drugs 16, 15 (2018). 12. Bartlett, T. C. Small scale experimental systems for coral research: Considerations, planning, and recommendations. NOAA Technical Memorandum NOS NCCOS 165 and CRCP 18, p. 68 (2013). 13. Farag, M. A. et al. Metabolomics reveals biotic and abiotic elicitor efects on the sof coral Sarcophyton ehrenbergi terpenoid content. Sci. Rep. 7, 648 (2017). 14. Mendes, C. et al. Efect of LEDs light spectrum on success of fragmentation and growth of leather coral Sarcophyton spp. Int. J. Aquacult. 7, 57–63 (2017). 15. Yuen, Y. S., Yamazaki, S. S., Nakamura, T., Tokuda, G. & Yamasaki, H. Efects of live rock on the reef-building coral Acropora digitifera cultured with high levels of nitrogenous compounds. Aquacult. Eng. 41, 35–43 (2009). 16. Bik, H. M. et al. Microbial community succession and nutrient cycling responses following perturbations of experimental saltwater aquaria. mSphere 4, e00043-19 (2019). 17. Costa, A. P. L. et al. Te efect of mixotrophy in the ex situ culture of the sof coral Sarcophytoncf. glaucum. Aquaculture 452, 151–159 (2016). 18. Khalesi, M. K., Beefink, H. H. & Wijfels, R. H. Flow-dependent growth in the zooxanthellate sof coral Sinularia fexibilis. J. Exp. Mar. Biol. Ecol. 351, 106–113 (2007). 19. Khalesi, M. K., Beefink, H. H. & Wijfels, R. H. Light-dependency of growth and secondary metabolite production in the captive zooxanthellate sof coral Sinularia fexibilis. Mar. Biotechnol. 11, 488–494 (2009). 20. Rocha, R. J. M., Calado, R., Cartaxana, P., Furtado, J. & Serôdio, J. Photobiology and growth of leather coral Sarcophyton cf. glaucum fragments stocked under low light in a recirculated system. Aquaculture 414, 235–242 (2013). 21. Rocha, R. J. M., Serôdio, J., Leal, M. C., Cartaxana, P. & Calado, R. Efect of light intensity on post-fragmentation photobiological performance of the sof coral Sinularia fexibilis. Aquaculture 388, 24–29 (2013). 22. Schutter, M., Kranenbarg, S., Wijfels, R. H., Verreth, J. & Osinga, R. Modifcation of light utilization for skeletal growth by water fow in the scleractinian coral Galaxea fascicularis. Mar. Biol. 158, 769–777 (2011). 23. Hellstrom, H. & Benzie, J. A. H. Robustness of size measurement in sof corals. Coral Reefs 30, 787–790 (2011). 24. Fabricius, K. E. Slow population turnover in the sof coral genera Sinularia and Sarcophyton on mid- and outer-shelf reefs of the Great Barrier Reef. Mar. Ecol. Prog. Ser. 126, 145–152 (1995). 25. Pupier, C. A., Bednarz, V. N. & Ferrier-Pagès, C. Studies with sof corals—recommendations on sample processing and normaliza- tion metrics. Front. Mar. Sci. 5, 348 (2018). 26. Sella, I. & Benayahu, Y. Rearing cuttings of the sof coral Sarcophyton glaucum (Octocorallia, Alcyonacea): Towards mass produc- tion in a closed seawater system. Aquacult. Res. 41, 1748–1758 (2010). 27. Rocha, R. J. M. et al. Development of a standardized modular system for experimental coral culture. J. World Aquacul. Soc. 46, 235–251 (2015). 28. Zoccola, D. et al. Te World Coral Conservatory (WCC): A Noah’s ark for corals to support survival of reef ecosystems. PLoS Biol. 18(9), e3000823 (2020). 29. Burmester, E. M. et al. Te impact of autotrophic versus heterotrophic nutritional pathways on colony health and wound recovery in corals. Ecol. Evol. 8, 10805–10816 (2018). 30. Meyle, J. Suture materials and suture techniques. Perio 3, 253–268 (2006). 31. Mayfeld, A., Fan, T. Y. & Chen, C. S. Physiological acclimation to elevated temperature in a reef-building coral from an upwelling environment. Coral Reefs 32, 909–921 (2013). 32. Siebeck, U., Marshall, N., Klüter, A. & Hoegh-Guldberg, O. Monitoring coral bleaching using a colour reference card. Coral Reefs 25, 453–460 (2006). We thank Zong-Min Ye for his comments on the culture systems. Tis study was funded by the Ministry of Sci- ence and Technology of Taiwan (MOST 107-2611-M-291-004).
| (2020) 10:20200 | Vol.:(0123456789) www.nature.com/scientificreports/
T.C.C. and T.Y.F. designed the study and performed the experiments. T.C.C., A.B.M., and T.Y.F. analyzed the data and wrote the manuscript.
All authors declare no confict of interest, be they fnancial or otherwise. Te authors declare that the research was conducted in the absence of any commercial or fnancial relationships that could be construed as a potential confict of interest. Supplementary information is available for this paper at https ://doi.org/10.1038/s4159 8-020-77071 -5. Correspondence and requests for materials should be addressed to T.-Y.F. Reprints and permissions information is available at www.nature.com/reprints. Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional afliations. Open Access Tis article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. Te images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.
© Te Author(s) 2020
| (2020) 10:20200 | Vol:.(1234567890)